What Drives Undocumented Immigration Policy, Economic, and Social Factors in the US and Mexico

Dana J. Smith is a PhD Candidate at the Dyson School of Applied Economics and Management at Cornell University.  Her research interests lie at the intersection of labor and development economics. She will be on the job market in Fall 2023.

Around ten million undocumented immigrants currently live in the United States, and around half these immigrants come from Mexico. The political rhetoric often focuses on place-specific policies, such as those that create “sanctuary cities” on the one hand, or that promote “self-deportation” on the other. Recent laws in Florida appear to already be pushing working immigrants out of the state. But how much do these state-specific laws impact the original decision to migrate to the US?

In my paper, I study the factors that impact undocumented migration flows from Mexico to the US. Using unique data, I am able to see where someone moved from, and what state in the US they went to. This allows me to analyze both push and pull factors simultaneously, while also taking into account pair-specific factors like the distance between states or previous migration networks. I consider drivers like wages and unemployment, the political environment, and violence. I focus on the years 2006 to 2013 which covers dramatic economic changes in both the US and Mexico due to the Great Recession, the introduction of more than 600 state-level immigration laws, and rising homicide rates in Mexico related to the government crackdown on cartels (the “decapitation” strategy).

I use data from the Mexican government on Matrículas Consulares, a form of ID card for Mexican nationals living abroad, to measure state-to-state, annual migration flows. Since undocumented immigrants lack other forms of identification, it is commonly assumed that most of these cards, which do not require someone to verify their residency status, go to undocumented immigrants. The Mexican government frequently advertises these ID cards to undocumented immigrants so the government can find and serve them if there is ever a crisis in the US and these are often the only government ID cards undocumented immigrants can get. Having a government ID in the US helps people participate in the local economy and community, for example by opening a bank account or even obtaining a library card. I collect all immigration laws from the National Conference of State Legislatures and create a policy index that captures the entire immigration-related legal environment in a state and year. States with more generous policies have higher, positive scores while states with more restrictive policies have lower, and often negative scores. I only include laws that can have significant impact on people’s lives, excluding policies like the Florida 2010 law that “honors Edith Lowngard Loebenberg, a Holocaust survivor who fled Germany for New York City” (NCSL). An example of a restrictive law would be the 2007 law in Oklahoma which “requires the verification of employment eligibility using the electronic employment verifications system (EEVS) and provides for a discrimination cause of action for the discharge of a US citizen while retaining an unauthorized immigrant on payroll,” commonly known as E-Verify (NCSL). Laws coded as “generous” included policies like allowing undocumented children to qualify for state public health insurance programs or eliminating visa requirements for driver’s licenses.

Identification for the results regarding economic conditions and homicides is based off the exogenous variation in these factors after accounting for fixed characteristics in the origin and destination states, as well as the state-to-state network effect. The policy index may be endogenous to the migration networks though as some US laws specifically mention the Matrícula Consular, such as those that allow someone to use this card to apply for a driver’s license. This might make people in certain states more likely to obtain an ID card, and thus appear in the data. I use an instrumental variables strategy to address this. I use state prison population per 100,000 residents to capture the strictness of the legal environment and find consistent results for both OLS (without the instrument) and 2SLS analysis.

I find that US conditions, or pull factors, tend to have a bigger impact on migration flows in general. In particular, high wages and low unemployment rates in US states increases migration into those states. This is in line with other work that shows many immigrants are economically motivated and also head towards areas where unemployment is low and jobs are more readily available. I find people leave states in Mexico where unemployment is higher. Higher homicide rates are also a key factor influencing migration patterns with higher outflows in states with higher homicide rates. Lastly, increases in the policy index, which would indicate a state becoming more generous or welcoming to migrants, has no impact on undocumented migration flows.

By simultaneously considering socio-political and economic conditions that may influence migrations, these results paint a more detailed picture of the factors that affect undocumented immigration from Mexico. Immigrants are economically motivated, but poor job opportunities alone do not explain the decision to migrant. I find people are also leaving violent areas, suggesting the line between “economic migrant” and “refugee” is not black and white. Homicide rates and disappearances continue to increase across Mexico, and criminal violence is frequently noted as part of the migration story for recent immigrants from Central America. Current state and federal immigration law does not always clearly address people fleeing criminal violence and my findings motivate giving more attention to updating current policies.  I also find that despite rhetorical claims, undocumented immigrants are not attracted to particular states because of their immigration policy. While these laws no doubt have important impacts on people’s lives once they live in the US, the initial decision to migrate does not depend on these local policies. This null result may reflect that even with strong networks people don’t know much or think much about the state policy environment before deciding to move but another likely interpretation is that the economy and violence are just the main motivators for people. E-verify may not deter someone who is fleeing cartel violence and the ability to get a driver’s license doesn’t attract someone who has a good job in Mexico.

Immigration is a huge part of American history and of all economies around the world. Understanding what drives people to move, and especially to undertake risky irregular migration, can help shape policy and prepare for future population flows. Future work may look at secondary migration decisions, for example, if undocumented immigrants leave Florida over recent policy changes, where do they go, and how does this impact the economy of Florida or the receiving states? Another important avenue for research regards the impact of climate change on migration. More local and higher frequency data, both quantitative and qualitative, will help us better understand modern immigration.

___________________________________________________

Interview with the author

Q: You use very cool data for this paper as well as your other research, what is your process of finding and getting access to these?

I first learned about this data in a really interesting paper on better ways to measure and then use migration flows from Mexico in economic analysis (Caballero, M. E., B. C. Cadena, and B. K. Kovak (2018). Measuring geographic migration patterns using matrículas consulares. Demography 55, 1119–1145.) The complete data, up through 2021 now, is available publicly from the Mexican government (https://ime.gob.mx/estadisticas/). They’ve really committed to sharing high quality data with researchers around the world. One thing I really appreciate about this data is that you can see both where people are from and where they go in the US, for the entire two countries. Other national data often only notes where someone is from and that they went to the US, or that an immigrant from Mexico is in the US but won’t indicate where in Mexico they are from. This data is actually available down to the municipal level in Mexico, which I am hoping to use for future project, but had to aggregate for this work because some of the other data wasn’t available at the municipal level. One downside is there is very little demographic data about the immigrants available publicly but given the sensitivity of the data that is completely understandable and it’s clear the Mexican government does a good job protecting people’s privacy.

Q: You don’t find a significant effect of destination state policies on the decision to migrate. Do you think this is because of lack of information on policies? Or do you think migrants prefer to move to places where they have connections already present?

I’ve spoken with a few people who immigrated to the US about this and I would say some of the effect may be due to not knowing too much about the legal environment, although policies that get a lot of attention in the US news are also often covered in Mexican and other Spanish-language news so I wouldn’t say people have no idea about state-level US immigration policy. Networks can also transmit a lot of this information. This data is a bit older so social media wasn’t as common a way to share news but now I see TikTok videos discussing Florida’s immigration policy! I think other factors are much more salient for people when they decide move to the US. While these policies definitely impact people’s lives they don’t necessarily influence the decision to migrate, which is a huge, costly event and is driven by a person’s livelihood and immediate safety. Once in the US someone may re-evaluate where they are based on changes in policy but that’s after they’ve undertaken the cost to just move to the US and I’m not particularly surprised that changes in policy aren’t spurring the initial migration decision. I think this is also a great opportunity for more qualitative researchers to explore and discuss the specific thought process behind migration decisions.